image
Name Grades

CRA's Solution for Common Name Procedures

Consumer Reporting Agencies (CRAs) that operate in the United States have a requirement to have procedures to assure maximum possible accuracy. In fact, The Consumer Financial Protection Bureau (CFPB) has taken enforcement actions against CRAs and mandated the following:

Revise their compliance procedures: The companies will revise procedures to assure reporting accuracy. These procedures include using algorithms to distinguish records by middle name and match common names and nicknames, using consumer dispute data to determine the root causes of errors, and using software to identify and reconcile discrepancies.
https://www.consumerfinance.gov/about-us/newsroom/cfpb-takes-action-against-two-of-the-largest-employment-background-screening-report-providers-for-serious-inaccuracies/

One procedure that is clearly recommended and necessary is a policy to treat common names differently.

  • How can that be done?
  • What constitutes a common name?
  • How can that be done with any level of objectivity?

Name Grades is the Solution

Name Grades is the solution for Consumer Reporting Agency common name procedures. Name Grades uses a proprietary algorithm to source through decades of registered names in an effort to weigh how common any given name is. This allows the CRA to establish a standardized, empirical policy to treat common names with a greater standard of care to ensure maximum possible accuracy.

How Name Grades Works

Data Required
  • First Name
  • Middle Name (optional)
  • Last Name
  • Gender (optional)
  • Year of Birth
Calculate Score

Name Grades' proprietary algorithm calculates a score for each element provided to get an aggregated commonality score between 1 to 100; 1 being an extremely uncommon name, and 100 being an extremely common name.

Deliverable

The Name Grades is delivered to you within our application or via an API call and return. The CRA can then determine for themselves, what score constitutes a common name and require additional measures to be taken with consumer reports resulting in a common name score.

Data Driven Results

Name Grades preprocesses historical data using a combination of machine learning and statistical methods to produce trillions of name combinations for each of the last 130 years. Name Grades then utilizes probabilistic models to estimate name probabilities. Name Grades's algorithm estimates automatically learn from name combinations and therefore a given score may shift slightly as the distribution changes.

Name Grades Examples

  • Extremely Common Name
  • Common Name
  • Uncommon Name
  • Extremely Uncommon Name
  • John Smith / Male / 1980

    Request
    
     {
    	"firstName": "John",
    	"gender": "M",
    	"lastName": "Smith",
    	"yearOfBirth": "1980"
     }
    											
    Response
    
     {
    	"score": 95
     }
    											
  • Allison Driver / Female / 1988

    Request
    
     {
    	"firstName": "Allison",
    	"gender": "F",
    	"lastName": "Driver",
    	"yearOfBirth": "1988"
     }
    											
    Response
    
     {
    	"score": 42
     }
    											
  • Hajidla Peterson / Female / 1995

    Request
    
     {
    	"firstName": "Hajidla",
    	"gender": "F",
    	"lastName": "Peterson",
    	"yearOfBirth": "1995"
     }
    											
    Response
    
     {
    	"score": 15
     }
    											
  • Divin Saju / Male / 1964

    Request
    
     {
    	"firstName": "Divin",
    	"gender": "M",
    	"lastName": "Saju",
    	"yearOfBirth": "1964"
     }
    											
    Response
    
     {
    	"score": 1
     }